Recent theories of drug abuse suggest that addiction involves aberrant memory processing. In nonhumans, stimulant drugs facilitate both encoding and retrieval processes, either of which could influence future drug-seeking behavior. In humans, subjective states of arousal or mood are known to bias emotional memory, raising the possibility that drugs producing these states may also bias memory and thereby influence drug-taking. The effects of drugs of abuse on memory have not been carefully studied in humans. In this exploratory study we propose to examine the effects of d-amphetamine (AMP) on encoding and retrieval processes, with a focus on emotional memory. We will address three main questions: 1) does AMP administered at the time of encoding facilitate subsequent retrieval assessed in a drug free state? 2) does AMP administered at the time of retrieval improve memory for stimuli previously encoded in a drug-free state? 3) is retrieval enhanced if the individual is in the same state (AMP or placebo) during both encoding and retrieval, compared to receiving AMP during either encoding or retrieval? The study will use a four-session within-subject design in healthy young adults (N=60). Each session consists of two phases, an encoding phase, followed 48 hours later by a retrieval phase. Subjects will receive capsules containing AMP (20 mg) or placebo before encoding or retrieval in all four combinations. During the encoding phases subjects will view or study standardized stimuli with emotionally positive, negative and neutral content, and their memory for the stimuli will be assessed 48 hours later during the retrieval phase. Based on findings with laboratory animals, we hypothesize that AMP administered during either encoding or retrieval will facilitate memory for stimuli with emotional content. Based on human studies of arousal and mood-congruence, we further hypothesize that memory facilitation may be related either to the drug's potential to increase either arousal or positive mood states. Finally, based on evidence for state-dependent learning, we hypothesize that memory will be maximized when encoding and retrieval take place in the same (drug or no-drug) state. The memory processes studied here have important consequence for drug-seeking. Drug-seeking behavior is strongly influenced by memory of previous events, and drug- induced biases during either encoding or retrieval could profoundly influence future drug use. This project will form the basis for future investigations of the effects of memory biases on drug-seeking behavior.